Test data: Achievement gaps persist
Across student groups’ races and backgrounds
Racial and need-based disparities in student achievement are not new problems in Connecticut schools. But state test data released last week was sobering proof that a global pandemic did no favors for closing those gaps.
For student groups who on average already struggled to meet proficiency standards, school closures and distance learning seemed to reverse a trend of improvement seen pre-pandemic.
Students considered to have “high needs” — a catchall used by the education department to include those learning English, with disabilities and from low-income backgrounds — saw their growth trend upward before the 2019-20 school year. Not only did they reflect their highest achievement on the 2018-19 tests, but also their results were improving at a faster rate than the state as a whole, according to reports at the time.
Then came COVID-19. Now, the latest data show a lower percentage of those same students were deemed proficient on the Smarter Balanced Assessment Consortium, or SBAC, testing in English language arts and math.
In fact, the test results suggest student achievement has suffered across the board, particularly in math, with the greatest impacts among hybrid and remote learners.
The education department report also found students of most races and ethnicities lost ground academically since the 2018-19 school year, with Black and Hispanic or Latino students averaging lower proficiency rates than their white and Asian peers — a trend that pre-dated the pandemic.
“Relatively large gaps like these don’t close in a couple years,” said Eric Loken, a UConn education professor affiliated with the Measurement, Evaluation and Assessment program. “And certainly they don’t close across a major disruption to the system.”
Race and ethnicity
The analysis only considered the scores of in-person test-takers, which made up 82 percent of third through eighth graders, though these students could be hybrid or remote learners. Department officials said the decision was made out of an abundance of caution, after finding some remote test scores higher than expectations.
“If we’re going to make state-level, high-level inferences, we’d be safer staying with test scores that come from tests that are administered in-person — in a setting, in a method, that we’ve done for years,” said Ajit Gopalakrishnan, the chief performance officer at the department.
More than half of white and Asian students met grade-level benchmarks in English, and more than a third did in math.
Among Black and Hispanic or Latino students, roughly a third or fewer in grades 3 through 8 were considered proficient readers, depending on learning model, the data show. About a quarter and below met grade standards in math.
Students who learned mostly in-person scored better than their peers in hybrid or mostly remote formats.
Asian students showed the smallest gap in growth across student learning models, the department found. Hybrid or remote learners who were Asian lost the least ground in English language arts compared to their peers.
High needs
A department analysis considered how the same students in grades 5 through 8 grew from one year to another, including would-be hybrid and remote learners. Third and fourth graders, who also sat for the tests this year, were not old enough to take the exams the last time they were administered.
Of the fifth to eighth graders, close to 36 percent of in-person learners with high needs were considered proficient in English language arts this year, compared 39 percent before the pandemic. Gaps were even greater for remote students: a quarter met grade-level standards, compared to a third pre-pandemic. More than three-quarters of in-person learners without high needs were proficient in English this year.
Gaps in math achievement, varying with learning models, were larger for fifth to eighth graders. Students learning in-person with high needs saw a proficiency rate of 22 percent, compared to 36 percent pre-pandemic. But of these remote learners — whose rates averaged above a quarter proficient in 2018-19 — just 10 percent met test benchmarks. Close to two-thirds of students learning in-person without high needs were considered proficient this year.
But missing from those percentages are the students who did not sit for the test in-person or at all.
Gopalakrishnan called this year “a level of nonparticipation we don’t normally see.”
The not-tested rate was greater among students with high needs, 21 percent, as compared to their nonhigh needs peers, 13 percent. Students with high needs also made up more than 70 percent of remote testers and not tested, begging the question if gaps in data could complicate assessing where all students are academically.
“If more remote learners had tested in person, potentially more students with high needs would have been part of that test population,” said Gopalakrishnan.
“While we didn’t have a state assessment measure on those students, we know that the districts are working with those students,” he said. “And we know that we’re doing what we can from our end to make sure they’re engaged and ready for this new school year.”
Officials at the education department said the state will continue to offer supports in response to the data, and asked that local education agencies, too, use the numbers to inform next steps.
“What we’re going to continue to ask districts to do is understand the learner,” said Irene Parisi, the chief academic officer at the department.
Parisi suggested that the state test results could be useful to help students progress in their studies, rather than fall back on what she called “vicious cycles of remediation.”
“Use these data, along with the data you have indistrict, and build out that profile of the learner,” she said. “Understand them, and match the resources and interventions and instruction to the need.”